Potential Diagnostic and Prognostic Values of Hematological Biomarkers in COVID-19‎ ‎

INTRODUCTION

The last few decades are proved to have been worst for human lives around the globe [1]. SARS-COV-2 is the third virus that causes fatal respiratory disease in humans. The recent deadly pandemic of a new respiratory virus called COVID-19, was first reported from the city of Wuhan, China.  Corona virus disease demonstrated its strong and rapid epidemic potential with spread worldwide in just a span of two months period [2].The World Health Organization (WHO) declared global emergency of COVID-19, due to rapid rise in cases of in China and many other countries by the mid of February 2020. By the date 20^th Feb 2020 the number of laboratory confirmed cases reached 60,000 in China with 1,700 deaths. ‎Likewise by the start of March 2020 virus had reached more than 30 countries outside China [3].‎Globally, since the 31st December 2019 and till date (28^th Nov 2020), a total of 61,715,119 cases have been reported worldwide, in accordance with the definition of the Centre of Disease Control (CDC) of COVID-19 cases, with 1,444,235 deaths [4]. By 29^th Nov 2020, the reported data from government sources in Pakistan declared 392,400 confirmed cases with 7942 deaths. Out of the total cases, 338000 cases are recovered so far and 1316 cases are under treatment in high dependency and intensive care units countrywide.  In Khyber Pakhtunkhwa, the total number of corona cases is crossing 46604 with 1355 deaths [5]. The gold standard test for diagnosis of COVID-19 is RT-PCR that detects SARS-COV-2 in high sensitivity and specificity limits and is used for symptomatic patients in acute phases [6].Increasing scientific research in field of COVID-19, the abnormalities in routine laboratory investigations, particularly the baseline investigation that is complete blood picture, has the potential to indicate the severity of the disease in a quick, easy, affordable, accessible and economical way and can give a way forward for the specific tests for diagnosis and prognosis of the disease [7]. Laboratory abnormalities in hematological indices are used for assessment of homeostasis, in order to know the impact of the COVID-19 on hematopoietic system and homeostasis which significantly suffer in this deadly disease [8].Present study was designed to assess the clinical sensitivity of the different indices of peripheral blood test (Complete blood count/CBC) to predict the severity and mortality in COVID-19.

MATERIALS AND METHODS

The reports of CBC of 182 cases admitted in the COVID-19 isolation unit in two tertiary care hospital of Khyber Pukhtunkhwa (Qazi Hussain Ahmed Medical Complex Nowshera & Hayatabad Medical Complex Peshawar) were analyzed from 1^st May 2020 to 30^th Aug 2020.

Ethical Approval was taken from Institutional Ethical Review Board of Nowshera Medical College Nowshera/Qazi Hussain Ahmed Medical Complex Nowshera via letter No- 1114-22/HD/QHAMC/NMC Dated 4^th April 2020 from the office of the Hospital Director. Verbal consent was taken from all the respondents, with confirmation that their confidentiality shall be maintained. Likewise the co-author obtained an ethical approval from HMC Peshawar on the research proposal of hematological and inflammatory markers and their impact on mortality in COVID-19 via Letter no 316/HEC/B&PSC/2020 on 15^th May 2020.

This study was conducted in line with the research regulations, followed the sound medical practice, redetected human rights and also within the principles of declaration of Helsinki of World Medical Association.

Patients from whom nasopharyngeal swab was obtained and thereafter hospitalized with an initial diagnosis of the corona virus disease irrespective of age and gender were included in the study.

Patient admitted in isolation with no reports of PCR and CBC were excluded from the study. The technique for the blood sampling for the CBC test was to collect 3 ml of venous blood by vein-puncture under aseptic techniques. The blood was added in an EDTA vacutainer at a concentration of 1.5mg/ml and mixed gently. Complete blood count was calculated on five parts Sysmex differential hematology analyzer.

For analysis purposes the patients were grouped as following to determine the statistical significance in dichotomous groups.

1. Patients discharged versus patients expired
2. Gender groups (Male and females)
3. Hb% <11g/dl and >11g/dl
4. TLC<11k and >11k
5. Platelet <150k and >150k
6. ANC<3.9 and >3.9k
7. ALC<1.5k and >1.5k
8. NLR<1.8 and >1.8

Data was entered in the SPSS (IBM Corp, Armonk, NY, USA) version 25 for analysis. The normality of data distribution was determined by Shapiro wilk test.

Peripheral blood information’s regarding Hb% (g/dl), TLC (x10.e3/ul), ANC (x10.e3/ul), ALC (x10.e3/ul), NLR, Platelet count were compared in groups (discharged satisfactorily vs expired).

Independent t test was used to determine the difference of the numerical variables. Non-parametric test i.e. Mann Whitney U Test was used where the data was not normally distributed.

Receiver operating characteristics (ROC) curve was used for estimation of the area under curve (AUC) to determine the relationship of clinical sensitivity of different hematological indicators and to predict the worst outcome in COVID-19. The sensitivity and specificity were calculated using diagnostic statistics with ROC curve and 95% confidence interval. (Table 1, figure 1).

Some of the indicators like neutrophilic-lymphocytic ratio (NLR) was calculated manually (N/L*100) from the CBC report.

RESULTS

In present study we determined the relationship of clinical sensitivity and specificity of different ‎hematological indicators to predict the worst outcome in COVID-19. A total of 182 patients, 135 ‎‎(74.2%) males and 47(25.8%) females were studied. ‎

Furthermore  NLR was main prognostic factor in COVID-19 to predict mortality/worst outcome ‎with an Area Under curve on ROC of 0.68 followed by absolute neutrophilic count (ANC) with ‎AUC value of 0.6 and Platelet count with AUC value of 0.55 on ROC. The AUC of the TLC was ‎at margin line, while for hemoglobin, it was not statistically significant. (Figure 1).‎

Based on the difference in the clinical sensitivity on ROC curve, we also determined the values of ‎the sensitivity, specificity in comparison with RT-PCR. The sensitivity pattern of different ‎hematological parameter as diagnostic marker in COVID-19 were: Hb%<11g/dl (59.1%), ‎TLC>11k (75.9%),  ANC>3.9% (86.2%), ALC1.8 (93.1%) and  ‎PLT<150k-(65.5%). (Table 1).‎

A statically significant higher values of in mean NLR and ANC values were noted in both study ‎groups i.e. deceased versus survivors (p=0.022, p=0.032 respectively). We also observed that the ‎patients who expired had low platelet counts as compared to the survivors (p<0.05).  (Table 2).‎

DISCUSSION

Considering the higher rate of infectivity and mortality due to COVID-19, an early detection of the ‎disease and understanding the prognostic values of the simple tests that are available even at ‎primary and secondary care centers, is the need of time. In present study we noted that the clinical ‎sensitivity of NLR was 93.1% with Area Under curve on ROC of 0.68. This documents it to be a ‎major prognostic factor in COVID-19 to predict mortality/worst outcome with a followed by ‎absolute neutrophilic count (ANC). A higher TLC, ANC and NLR carry bad prognosis in case of ‎COVID-19. No doubt the definite diagnosis of the COVID-19 is by RT-PCR. Asymptomatic ‎‎/subclinical patients deteriorate if not properly observed and looked after7. We analyzed the role of ‎complete blood count (CBC) and assessed various hematological parameters for their prognostic ‎values in groups with survivors against those met the unfortunate end of lives with COVID-19. We ‎observed that no significant difference in concentration of hemoglobin in both the groups with ‎different outcome (p>0.05). However there was a significant difference in indicators i.e. TLC, ‎ANC, ALC, NLR and Platelet counts among the survivors versus non survivors (p<0.05). Similar ‎findings have been reported from a study conducted in Ankara Turkey [9]. Another study from ‎Karachi Pakistan reported 364 Covid-19 patients, with 27.7% mortality. They assessed likewise ‎patients in survivors and non-survivor groups. They reported that there was a significant difference ‎in TLC (p<0.001), ANC (p<0.001) and ALC (p<0.001) [10].  Their findings were in concordance ‎with our results.‎

Another retrospective study by the same principal author, from different teaching hospitals of ‎Karachi, reported a significant difference of hematological markers in groups survived versus ‎expired. They reported a significant difference in mean Hb% (p-0.03), TLC, ANC, ALC and NLR ‎‎(p-0.001 each) [11]. Again their findings do match our results for WBC and platelet but we could ‎not find a significant difference in mean Hb% of the survivors versus deceased.‎

Other researcher findings do support the fact that NLR is diagnostic and prognostic marker in viral ‎infections (AUC=0.710) [12]. Ai-Ping Yang et al reported NLR with AUC of 0.743 on ROC to ‎determine the outcome of the seriously ill patients due to COVID-19 to document the prognostic ‎values of NLR [13] (Table 3).  Their findings closely match our results. Moreover the Sun et al ‎reported the AUC of 0.88 for NLR that was recorded in seriously ill COVID-19 patients who met ‎the worst outcome [14].‎

Likewise the AUC of the absolute neutrophilic count (ANC) was 0.6 and was significantly higher ‎in group with case fatality. Therefore ANC can also be used as prognostic marker in COVID-19. ‎Notably, a study published in American Journal of Hematology reported high leukocyte count in ‎COVID-19 ICU patients with severe complications with median peak ANC of 11600/cmm3 as ‎compared to non-ICU patients without complications (p<0.001) [15].‎

Platelet count itself is a marker in COVID mortality. We observed an AUC of 0.55 for low platelet ‎count in patient died of COVID-19. Our findings are in concordance with findings of the Yang et ‎al who also reported thrombocytopenia in all deceased patients due to Covid-19 in their target ‎population [16]. The comparison of the finding with other national and international studies is well ‎documented in table 3.‎ 

CONCLUSION